Muhammed Jaabir Mohamed Zifri

• Developing a machine learning model to identify gait activities from sparsely annotated accelerometer data (Signal data), with the goal of automating data labeling to improve analysis and model accuracy.

• Developed and deployed four AI-powered projects in four weeks using React JS, Next.js, Firebase, Clerk, Gen AI APIs, and Vercel, following agile methodologies and incorporating CI/CD practices.

• Collaborated with a team of three to build an interactive AI agent, "Rate My Teach" / "U-Rankly" integrating a custom RAG pipeline with OpenAI and Pinecone for user query responses.

• Worked with my team to design and develop a scalable full-stack AI web application, achieving over 100 sign-ups on the waitlist within three days of launch.

• Helped some startup companies with their backlogs by contacting them through social media, email and referrals.

• Provided ML & DL solutions for diverse domains like predictive maintenance, LLMs (Large Language Models) like text summarizer ( e.g. GPT models from Hugging Face ), Custom NER model.

• Successfully researched and implemented a chatbot for healthcare using a knowledge graph (Neo4j) and was invited as a speaker to IIT to talk about this project.

• Developed varieties of classification tasks involving models from simple statistical models to Computer vision tasks like detecting American Sign Language (ASL) and deployed the trained models to an online model hub ( e.g. Hugging Face).

• Led a team of 5 for several projects to develop and deploy multiple ML models into production web applications, meeting all defined KPIs set by the company’s CEO.